Definition:

Descriptive analytics is the set of techniques used to summarize and present observed data. It helps answer questions about what happened, how much, when, where, and how often within a defined period and scope.

Its results may be expressed through counts, percentages, averages, distributions, changes, tables, and charts. It describes the available information, but it does not automatically identify causes or predict what will happen next.

It can be applied to historical or recent data. The distinction does not depend on the age of the data, but on the question: showing yesterday’s sales and summarizing visits from the last five minutes are both descriptive uses when they are limited to representing what was observed.

How Descriptive Analytics Is Performed

The analysis begins with a specific question and a shared definition of the metrics. A summary can only be interpreted when its population, period, and calculation rules are known.

A common process includes these five stages:

Define the objective: establish which phenomenon is to be described and which decision or monitoring activity the result will support.
Define the data scope: specify sources, population, period, unit of analysis, segments, and inclusion or exclusion criteria.
Check quality: review missing values, duplicates, instrumentation errors, definition changes, and possible coverage biases.
Calculate summaries: obtain counts, proportions, measures of central tendency, dispersion, changes, or distributions suited to the question.
Present and interpret: use tables, charts, or a dashboard and document the context and limitations needed to read them.

The mean does not always represent a dataset well. The median, percentiles, and distribution may be more informative when there are extreme values or groups with different behaviors. Likewise, a percentage change needs its reference value and absolute volume to avoid disproportionate interpretations.

Comparison also requires equivalent bases. Changing the time window, the definition of a user, or the way an event is recorded can produce an apparent difference that does not correspond to a real change in the phenomenon.

Applications of Descriptive Analytics

Descriptive analytics is used in operations, finance, marketing, product, and customer service. Its purpose is to turn scattered observations into an understandable representation of the state and development of an activity.

Common applications include:

Activity monitoring: summarize orders, revenue, inventory, incidents, or service usage during a period.
Time comparison: observe daily, monthly, or year-over-year changes using compatible calendars and criteria.
Segment analysis: compare results by product, channel, device, location, or customer type without confusing an observed difference with its cause.
Pattern and exception detection: locate concentrations, seasonality, outliers, or deviations that require further investigation.
Communication of results: present KPIs and metrics with definitions, sources, and references that support interpretation.

In web analytics, for example, it can describe users, sessions, page views, events, traffic sources, and conversions. A decrease or increase shows what changed in the recorded data, but it does not by itself demonstrate whether the cause was a campaign, a technical incident, seasonality, or a measurement change.

Descriptive reports can guide new questions and help prioritize reviews. They do not automatically determine which action should be taken or guarantee that the observed pattern will continue.

Differences from Other Types of Analytics

Descriptive analytics is often presented alongside other approaches. The boundaries are not always rigid, but each primarily answers a different question:

Descriptive: summarizes what happened or the observed state.
Diagnostic: investigates why it may have happened through comparisons, breakdowns, and hypothesis testing.
Predictive: estimates unknown or future outcomes using models and assumptions, as explained by predictive analytics.
Prescriptive: evaluates alternatives and proposes actions according to objectives, constraints, and estimated consequences.

The same project may combine these approaches. A report describes a decrease in conversions; diagnostic analysis studies explanations; a model estimates future scenarios; and a prescriptive evaluation compares possible responses. Moving from one stage to another requires additional data, methods, and validation.

An observed correlation between two variables does not prove causation either. Experiments, quasi-experimental designs, or other evidence may be needed to attribute an effect. The value of descriptive analytics depends on representing the data faithfully and not claiming more than the data can support.